Abstract
In almond, the presence of double seeded nuts significantly reduces their commercial value. For this reason, in breeding programs of this species, seedlings producing double seeds are usually eliminated. In order to avoid this trait in the offspring and increase the efficiency of the breeding programs it is important to know its genetic control. In this work, heritability of double seeds was estimated by midparent–offspring regression (narrow sense, h2) and by variance components analysis (broad sense, H2) for 3 years, in a population with a combination of full-sib and half-sib offspring coming from eight crosses designed for this objective. Heritability was high by both methods every year, and the values calculated by variance components were higher than those estimated by regression. The average ratio h2/H2 0.69 implies some influence of a non-additive effect for this trait, such as dominance and environmental effect. In this case, the influence of the temperatures preceding blooming on the percentage of double seeds can be inferred from our results, with lower temperatures increasing the percentage of double seeds. In the crosses between parents with high percentage of double seeds a normal (Gaussian) phenotypic distribution was observed. For crosses between one parent without double seeds and parents with high-medium percentage of double seeds a right-skew distribution (L-shaped) was observed, and certain transgressive segregation was observed in the descendants. This transgressive segregation was confirmed when both parents presented null percentage of double seeds. The results confirmed the complex architecture of this trait transmitted quantitatively, where additive, non-additive and genotype-by-environment interaction effects play an important role